
SK hynix has started providing samples of its next-generation HBM4E DRAM to major customers, marking a key step in advancing memory solutions for artificial intelligence workloads. The company aims to accelerate commercialization by collaborating closely with partners to move toward mass production.
The new HBM4E chips are designed to meet the growing demands of AI-driven data centers and large-scale computing environments, where high bandwidth and efficiency are critical.
“SK hynix has laid the foundation to strengthen its AI leadership with HBM4E based on its market-leading technological capabilities and manufacturing expertise,” said Ahn Hyun, President and Chief Development Officer. In addition, Ahn Hyun said, “Through close collaboration with our partners, we will deliver the value needed in the market while reinforcing our technology leadership as a full-stack AI memory creator.”

The 12-layer HBM4E product delivers notable improvements in speed and energy use. It offers data processing speeds of up to 16Gbps per pin, alongside a power efficiency gain exceeding 20 percent compared to earlier models. These enhancements are expected to strengthen performance in AI training and inference tasks, where rapid data transfer and reduced energy consumption play a central role.
In addition, the memory solution reduces latency through an updated interface and optimized design, improving stability in high-bandwidth environments and enabling smoother data processing.
SK hynix has integrated its advanced MR-MUF (Mass Reflow Molded Underfill) technology into the HBM4E design. This manufacturing process enhances structural stability by protecting stacked chips with injected protective material.
The 12-stack configuration achieves a capacity of 48GB while maintaining durability. Heat resistance has also improved by 17 percent over the previous HBM4 generation, supporting reliable operation in demanding high-performance computing scenarios.
The company’s latest offering builds on its track record of delivering HBM3, HBM3E, and HBM4 solutions. By combining production expertise with proven reliability, SK hynix aims to address bottlenecks in AI infrastructure and support next-generation computing systems.
The introduction of HBM4E positions the company to strengthen its role in the expanding AI memory market, where demand for scalable, high-performance solutions continues to rise.
SK hynix leadership emphasized continued cooperation with industry partners to ensure timely production and deployment. The company believes its technology and manufacturing capabilities will reinforce its standing as a comprehensive provider of AI memory solutions.
As AI adoption accelerates globally, the launch of HBM4E reflects intensifying competition among semiconductor players to deliver faster, more efficient memory technologies for data-intensive applications.
20 June 2026